Thinking Machines launches Inkling 2 Small: 594B MoE scoring 31 on AA Index, Apache 2.0 weights next week
MohitIyyer · x · 2026-11-06
Thinking Machines Lab released its second-generation model Inkling 2 Small, initially available on Tinker.
- MoE with 594B total / 17B active parameters; supports image inputs and variable effort.
- Scores 31 on the Artificial Analysis Intelligence Index, a big jump over gen-1, driven by more efficient pre-training and a large unified RL run.
- Strong long-horizon agentic and vision skills: demos include a fully procedural Three.js city simulator and a from-scratch Python PC emulator that boots a real Linux kernel.
- Leads visual benchmarks vs GLM-5.3-Flash, DeepSeek-V4.1-Flash and Kimi K3 (MMMU Pro 81.6%, ScreenSpot Pro 83.8/91.5%, AA-3D 1021).
- Closed weights via Tinker now; full Apache 2.0 weight release planned next week after external safety testing and red-teaming.
Related event: Thinking Machines Releases Inkling 2 Small Open-Weight Model(42 posts)→
More from Models
- Ampersand hits 92.2% on SWE-bench Pro v2, 70% cheaper via human-AI co-coding — gharik · 2026-11-06
- AI Cyber Skill Doubles Every 4-5 Months, Task Horizon Grows 52x in a Year — FinanceYF5 · 2026-11-06
- 50 insiders vote on AI data companies shaping frontier research: Surge leads with 16 votes — FinanceYF5 · 2026-11-06
- OpenAI publishes two more misalignment incident reports: grader prompt injection and sandbox network bypass — scaling01 · 2026-11-06
- Thinking Machines releases Inkling 2 Small; RL training taught model to circumvent guardrails — johnschulman2 · 2026-11-06
- Qwen3.8-Flash-Next FP4 quantized model trends on Hugging Face — autotrust · 2026-11-06